AI Chips news and analysis on Web Pulse: 6914 curated articles tracking the latest AI Chips developments, tools, and research, updated continuously from vetted sources.
SemiAnalysis reported that Nvidia's Vera Rubin NVL72 platform delivered up to 7x better token throughput per megawatt than Blackwell on its AgentX agentic inference benchmark, even on early pre-release software, and esti…
Cornelis Networks raised $205 million to fund its expansion into scale-up networking and introduced Active Compute Fabric, an architecture that embeds programmable compute — including licensed third-party RISC-V cores — …
Nvidia CEO Jensen Huang took a surprise onstage call from President Trump at the All-In Summit in Los Angeles on Monday, where Trump told Huang "the robots will not be taking over" and called warnings about catastrophic …
Fujitsu announced on September 14 that it will begin global sales of FUJITSU-MONAKA, a 144-core 2-nanometer Arm server CPU, in November 2026, targeting AI inference workloads for banks, telecoms and governments seeking a…
MSI's XpertStation WS300 keeps NVIDIA's 1,300W GB300 Grace Blackwell Ultra Superchip from throttling in a desktop tower because MSI rates its liquid-cooling loop for 1,400W, roughly 100W above the Superchip's power limit…
Fujitsu will put its 2nm FUJITSU-MONAKA server family, built around a 144-core processor running at up to 3.8GHz, on sale in November for air-cooled AI inference and AI agent workloads. Fujitsu rates MONAKA at twice the …
Cornelis unveiled its Active Compute Fabric, a networking technology the company says is designed to maximize AI rack utilization by evolving the network from a passive transport layer into an active participant in compu…
FS packaged its AI optics portfolio into two parts: Scale-Out transceivers at 400G, 800G, and 1.6T for intra-cluster GPU links, and Scale-Across coherent modules at 400G and 800G reaching up to 500 km for data center int…
China's Ministry of Foreign Affairs spokesperson Guo Jiakun on Monday criticized Anthropic CEO Dario Amodei's call for the U.S. to curb China's AI capabilities, saying "fearmongering, confrontation and vicious competitio…
Apple asked Judge Edward Davila in federal court on Monday for accelerated discovery into OpenAI's internal communications, devices, accounts, and employees ahead of an October 1 injunction hearing, seeking evidence of w…
NVIDIA Transformer Engine is now integrated with JAX to accelerate dropless Mixture of Experts (MoE) training, targeting the conditional computation bottleneck in architectures used by DeepSeek, Qwen, and Mixtral. The in…
NVIDIA reported that combining its Transformer Engine library with the JAX Python library raised DeepSeek-V3 MoE training throughput on NVIDIA GB200 GPUs from an unoptimized baseline of 103 TFLOPS/GPU to 1,068 TFLOPS/GPU…
Arm introduced the Neoverse CSS N4, its most configurable Compute Subsystem to date, supporting up to 128 processor cores on a die plus LPDDR6 memory and PCIe Gen 7, and said the platform delivers up to twice the perform…
Cognichip unveiled Artificial Chip Intelligence (ACI), a physics-informed AI system for chip design that founder and CEO Faraj Aalaei says compresses months of front-end work into days, according to a September 14 Forbes…
Google announced general availability of Pause/Resume for Dataflow batch jobs and support for G4 VMs powered by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. Pause/Resume lets Dataflow customers resume failed long-r…
Anthropic CEO Dario Amodei published an essay on September 12 titled "We Must Pace the Frontier," urging the AI industry to slow development of advanced capabilities while Anthropic maintains a $517 billion compute commi…
Pinterest expanded its partnership with Nvidia, pairing Nvidia's Blackwell GPUs and Dynamo inference framework with Pinterest's proprietary visual embeddings to achieve an 85x improvement in response startup times and a …
Apple's first 2nm iPhone places the company at the front of a new chip race, according to Forbes contributor Jon Markman, with TSMC supply, rising costs and on-device AI raising the stakes. The report ties the 2nm chip t…
Crusoe opened its second manufacturing facility in Tulsa, Oklahoma, adding 400,000 square feet and bringing the AI factory company's national manufacturing footprint to roughly 1 million square feet. The two Tulsa sites …
Latham & Watkins, the second largest US law firm, is buying Nvidia hardware to fine-tune open-weight models in-house, according to a post on X by Anand Iyer. Iyer framed the move as an "enterprise sovereign AI stack" com…